Comparison of a novel chemiluminescent based algorithm to three algorithmic approaches for the laboratory diagnosis of Clostridium difficile infection

J. Goret, J. Blanchi, C. Eckert, S. Lacome, A. Petit, F. Barbut, C. Bébéar, Francis Mégraud

Rapid commercial assays, including nucleic acid amplification tests and immunoassays for Clostridium. difficile toxins, have replaced the use of older assays. They are included in a two-step algorithm diagnosis, including first the detection of the glutamate dehydrogenase (GDH) as a screening method and second the detection of toxins as a confirmatory method. Although assays that detect the presence of free toxins in feces are known to lack sensitivity, they are preferable to confirm infection. We evaluated the accuracy of the chemiluminescence-based method detecting C. difficile GDH and free toxins A/B (DiaSorin algorithm) to an enzyme-immunoassay (EIA) for GDH with a molecular toxins test (Meridian algorithm), EIA-GDH and an EIA-toxins A/B algorithm (Alere algorithm) with and without toxigenic culture for confirmation. A total of 468 diarrhoeal and loose stool samples were included in the study. A positive result was defined by a positive GDH and a positive toxin test. Discordant samples were resolved using an enriched toxigenic culture considered as the reference method. After resolution, the DiaSorin algorithm showed a high sensitivity (86.7 %) compared to that of the Alere algorithm with (60.0 %) and without (50.0 %) confirmation by culture and was as sensitive as the Meridian algorithm (90.0 %), while the specificities were similar: 99.1, 99.5, 99.5 and 98.9 %, respectively. The DiaSorin algorithm was as sensitive as an algorithm including nucleic acid amplification test for toxins. Chemiluminescence toxin-enhanced signal assay compensates the lack of sensitivity usually observed for EIA tests for toxins.